AR245-NE28-07 ARI 16 March 2005 1:3 V I E E W R S First published online as a Review in Advance on March 17, 2005 I E N C N A D V A Structure and Function of Visual Area MT Richard T. Born1 and David C. Bradley2 1Department of Neurobiology, Harvard Medical School, Boston, Massachusetts 02115-5701; email: [email protected] 2Department of Psychology, University of Chicago, Chicago, Illinois 60637; email: [email protected] Annu. Rev. Neurosci. Key Words 2005. 28:157–89 extrastriate, motion perception, center-surround antagonism, doi: 10.1146/ magnocellular, structure-from-motion, aperture problem by HARVARD COLLEGE on 04/14/05. For personal use only. annurev.neuro.26.041002.131052 Copyright c 2005 by Abstract Annual Reviews. All rights reserved The small visual area known as MT or V5 has played a major role in 0147-006X/05/0721- our understanding of the primate cerebral cortex. This area has been 0157$20.00 historically important in the concept of cortical processing streams and the idea that different visual areas constitute highly specialized Annu. Rev. Neurosci. 0.0:${article.fPage}-${article.lPage}. Downloaded from arjournals.annualreviews.org representations of visual information. MT has also proven to be a fer- tile culture dish—full of direction- and disparity-selective neurons— exploited by many labs to study the neural circuits underlying com- putations of motion and depth and to examine the relationship be- tween neural activity and perception. Here we attempt a synthetic overview of the rich literature on MT with the goal of answering the question, What does MT do? www.annualreviews.org · Structure and Function of Area MT 157 AR245-NE28-07 ARI 16 March 2005 1:3 pathway. Finally, MT neurons are concerned Contents with visual motion, which is of obvious etho- logical importance, which has been exten- INTRODUCTION ..................158 sively characterized psychophysically, and for MT WAS A KEY PART OF THE which there are well-defined mathematical EARLY EXPLORATION OF descriptions. Much of the work on MT has EXTRASTRIATE CORTEX ......158 focused on its role in visual motion process- CONNECTIONS....................159 ing, though, as we hope to make clear in what FUNCTIONAL ORGANIZATION . 162 follows, MT plays a richer and more varied BASIC TUNING PROPERTIES . 164 role in vision. SURROUND MECHANISMS .......165 THE COMPUTATION OF VELOCITY.......................168 MT WAS A KEY PART OF THE NOISE REDUCTION ...............173 EARLY EXPLORATION OF SEGMENTATION ..................173 EXTRASTRIATE CORTEX THE COMPUTATION OF Part of MT’s significance is historical; it STRUCTURE ....................174 played an important role in the discovery of EXTRARETINAL EFFECTS ........176 new extrastriate visual areas (Felleman & Van PERCEPTUAL CORRELATES Essen 1991) and in the idea that they consti- AND POPULATION CODES . 177 tute specialized representations of the visual Single-Neuron Sensitivity ..........177 world (Zeki 1978, Barlow 1986). Vector Summation versus At the beginning of the twentieth cen- Winner-Take-All ...............178 tury, primate visual cortex was thought to Distributed Speed and Acceleration consist of only three architectonically distinct Codes ..........................179 fields (Brodmann 1909). Beginning in the late CONCLUSIONS ....................179 1940s, however, it became clear that consid- erably more of the cortex was involved in vision. The first demonstration came from temporal lobe lesions that produced visual im- INTRODUCTION pairment (Mishkin 1954, Mishkin & Pribram The middle temporal visual area (MT or V5) 1954) unaccompanied by deficits in other sen- of the macaque monkey possesses a number of sory modalities (Weiskrantz & Mishkin 1958, attributes that have made it particularly attrac- Brown 1963). Mapping studies using sur- by HARVARD COLLEGE on 04/14/05. For personal use only. tive to systems neuroscientists. This region is face electrodes also revealed visually respon- typical of extrastriate cortex but is still readily sive regions well anterior to those tradition- identifiable both anatomically and function- ally associated with vision (Talbot& Marshall ally. Though extrastriate, it is still quite close 1941, Clare & Bishop 1954, Woolsey et al. to the retina—its principle inputs as few as five 1955). In addition, new anatomical techniques synapses from the photoreceptors—a feature (Nauta & Gygax 1954) permitted the label- Annu. Rev. Neurosci. 0.0:${article.fPage}-${article.lPage}. Downloaded from arjournals.annualreviews.org which means, among other things, that the ing of connections after lesions of striate cor- mechanisms by which its receptive field prop- tex (Kuypers et al. 1965, Cragg & Ainsworth erties arise can be profitably studied. And, al- 1969, Zeki 1969), which revealed a direct stri- though MT neurons are near enough to the ate (V1) projection zone situated on the pos- inputs to be mechanistically tractable, they are terior bank of the superior temporal sulcus also close enough to some outputs—in par- (STS). ticular, those involved in eye movements— MT was discovered at roughly the same to provide an easily measurable, continuous time by two different groups. In England, readout of computations performed in this Dubner & Zeki (1971) were able to record 158 Born · Bradley AR245-NE28-07 ARI 16 March 2005 1:3 Figure 1 First demonstration of direction selectivity in macaque MT/V5 by Dubner & Zeki (1971). (a)Neuronal responses to a bar of light swept across the receptive field in different directions (modified from figure 1 of Dubner & Zeki 1971). Each trace shows the spiking activity of the neuron as the bar was swept in the direction indicated by the arrow. The neuron’s preferred direction was up and to the right. (b) Oblique penetration through MT (modified from Figure 3 of Dubner & Zeki 1971) showing the shifts in preferred direction indicative of the direction columns subsequently demonstrated by Albright et al. (1984). See also Figure 4. visual responses from the V1-projection zone Following the first studies, a series of pa- in anesthetized macaques, in so doing es- pers confirmed that MT contained a high tablishing a number of physiological hall- concentration of direction-selective neurons marks, particularly their direction-selective in several species of both New and Old preferred direction: the direction of responses (Figure 1a). Quite presciently, they World monkeys (Zeki 1974, 1980; Baker et al. motion eliciting the also suggested a columnar organization for 1981; Van Essen et al. 1981; Maunsell & Van greatest response direction-selective neurons (Figure 1b) and Essen 1983a,b; Felleman & Kaas 1984). These from a given neuron a role for MT signals in guiding pursuit studies indicated that MT was both unique eye movements, both subsequently confirmed as a cortical area highly specialized for visual (Albright et al. 1984, Lisberger et al. 1987). motion and, at the same time, common to a Around the same time, Allman & Kaas (1971) number of different primate species. were recording from owl monkeys and using a different approach. They made systematic rows of microelectrode penetrations across CONNECTIONS the entire cortex, mapping receptive fields as Like every other cortical area, MT has a rich by HARVARD COLLEGE on 04/14/05. For personal use only. they went, thus discovering a large number of set of interconnections with other regions of retinotopically organized maps. One of these, the cortex as well as with numerous subcor- which they named MT for middle temporal, tical structures. These connections have been mapped onto a well-defined region of dense discussed in previous publications (Felleman myelination in the lower layers and contained &Van Essen 1991, Orban 1997, Lewis & neurons that responded better to drifting bars Van Essen 2000), so we do not recapitulate Annu. Rev. Neurosci. 0.0:${article.fPage}-${article.lPage}. Downloaded from arjournals.annualreviews.org than to flashed spots. The myelination was them here. From a broad perspective, MT’s also later shown to be characteristic of the corticocortical connections identify it as one macaque motion area (Van Essen et al. 1981), of the main inputs into the dorsal or poste- which Zeki subsequently named V5. This his- rior parietal processing stream (Ungerleider tochemical feature has been an underappreci- & Mishkin 1982, Maunsell & Newsome ated factor in contributing to the detail with 1987), and its key outputs target structures which MT has been studied because it has per- that are implicated in the analysis of optic flow mitted reliable comparisons across different (e.g., MST, VIP) and the generation of eye studies. movements (e.g., LIP, FEF, SC, dorsolateral www.annualreviews.org · Structure and Function of Area MT 159 AR245-NE28-07 ARI 16 March 2005 1:3 Figure 2 Gestalt map of major routes into MT in the manner of Felleman & Van Essen (1991). Line thickness is roughly proportional to the magnitude of the inputs, on the basis of a combination of projection neuron numbers and, where data are available, the characteristics of their axon terminals (see Figure 3). The thickest lines represent the direct cortical pathway emphasized in the text. Following are important caveats: The pathways shown are those discussed in the text and omit a number of known feedforward cortical inputs that appear lesser in magnitude (V3A, VP, PIP) as well as many subcortical inputs. The sources of the direct and indirect projections from V1 are probably not defined purely by cell morphology (i.e., spiny stellate versus pyramidal; see Elston & Rosa 1997), though they are largely distinct; the largest 4B cells contribute to the direct pathway (Sincich & Horton 2003). The precise nature of the retinal inputs to K1,2 is not known, though their response properties are W-like in the galago (Irvin et al. 1986). Also, the proposed input to MT from the SC via the pulvinar is rendered problematic by the finding that, in owl monkey pulvinar, the principle target of SC terminals (PICM)is by HARVARD COLLEGE on 04/14/05. For personal use only. different from the main source of MT projections (PIM) (see Stepniewska et al.